Towards a Knowledge-driven Maintenance Support System for Manufacturing Lines.

This paper presents how to design a Knowledge-driven Maintenance Support System (MSS) to prognostic breakdowns in production lines and how it affects to the production rate. The system is based on the subcycle time monitorization and how the cycle time variability of machine parts can be used as a deterioration indicator that could describe the dynamic of the failure for the machine parts. For this proposal, a novel model based on mini-terms and micro-terms introduced in our previous work as a machine subdivision is used. A mini-term subdivision can be selected by the expert team for several reasons, the replacement of a machine part or simply to analyze the machine more adequately. (A micro-term is a component from a mini-term and it can be as small as the user wishes. Without loss of generality, the paper focuses its attention on a welding line at Ford Motor Company located at Almusafes (Valencia) where a welding unit was isolated and tested for some particular pathologies. The cycle time of each mini-term is measured by changing the deteriorated components in the cycle time. The deterioration of the parts (a proportional valve, a cylinder, an electrical transformer, the robot speed and the loss of pressure) are tested within the range of normal production, which is the range that cannot be detected by alarms or maintenance workers but when the change point is occured. The statistical analysis of the data obtained in the experiments allows us to define the rules that govern the decisions for the real-time Knowledge-driven MSS. This analysis and the welding line simulation also allows us to know the loss of productivity when the change point occurs. In the worst case, the welding line reduces their production rate almost 40%.